Role Autonomous Decentralized System architecture towards system evolution

During the past two decades, autonomous decentralized system (ADS) has resolved online extensibility, online maintenance and fault tolerance of system, and realized dynamic extension of system in node level and module level. However, with the developing of research on ADS, autonomously expression of system function is needed for achieving evolutional ability in function level. In this paper, we analyze the temporal and spatial expression mode of system function, and abstract a new kind of application mode which autonomously expresses latent function of system. Based on the analysis, the new design and development concept of system such as separating strategy and mechanism; separating function and its expression; separating dynamic and static data, and team partaking to function are proposed. And a role field and data field dual driven architecture (RDDA) model is proposed, including role field, data field, role autonomous processor, data arbitrator, role-function-object dispatcher etc. One design instance of RDDA is presented in the last part of this paper, whose kernel is configuration file and role autonomous processor. And the possible application situation is also discussed. RDDA achieves the unified description of master-slave structure, client-server structure and peer to peer structure. The structure of RDDA support temporal and spatial dynamic evolution on the level of function expression (role level), which can improve the system ability of rapid rebuilding and dynamic reconfiguration.